irclog2html for #brlcad on 20060214

01:55.21 Twingy grizzly nailer is much better *grunt*
02:56.02 Twingy Live is horribly gay band
02:56.24 Twingy they should change their name to Xandir
02:56.43 Twingy or.. Live action squirrel with big balls
03:08.15 brlcad heh
04:07.31 *** join/#brlcad mahesh (n=mahesh@12-217-229-38.client.mchsi.com)
04:08.11 mahesh Sean, had a question for you
04:15.35 brlcad mahesh: fire away
04:49.38 *** join/#brlcad mahesh (n=mahesh@12-217-229-38.client.mchsi.com)
04:50.32 mahesh Sean, you there?
04:52.16 brlcad still here :)
04:53.23 mahesh oh there you are
04:54.00 mahesh had a question for you
04:54.16 brlcad i'm always here in some form or another, best to just ask whatever question regardless :) i'll answer when you/I return
04:54.29 mahesh ok
04:55.45 mahesh the remote nodes after raytracing, return the application structure to the master node
04:56.03 mahesh i am trying to do a view_pixel on that application structure
04:56.25 mahesh but it is not showing up on the framebuffer
04:57.01 mahesh i want to know, if the remote nodes have actually done the work (raytrace)
04:57.16 mahesh is there a way to find out from any of the fields in that structure?
04:57.26 brlcad a_color should be set
04:57.29 mahesh like a_color or something?
04:57.57 mahesh what could be the values of a_color[]?
04:58.30 brlcad 1/1/1 if it "missed", something else otherwise
04:59.34 mahesh ok....lemme print and check
05:02.13 brlcad might be useful to raytrace a tiny picture like -s2 so you can test individual values
05:03.05 brlcad fi, right-clicking the framebuffer will report the pixel value
05:04.38 mahesh hmm...interesting. i am printing out individual values and they are all like 0.0000, 0.00000.0.39222
05:05.08 mahesh correction 0.00000, 0.00000, 0.003922
05:05.53 brlcad er, they are chars for starters :)
05:06.00 brlcad not floats
05:06.19 mahesh oh i tried without -s2 option
05:07.54 brlcad and I misspoke, 0/0/1 if it misses
05:08.05 brlcad which is what that value looks like
05:08.23 brlcad s2 is just to set the image size to 2x2
05:08.29 brlcad instead of 512x512
05:08.53 mahesh so, there is definitely something wrong on the remote node end right?
05:09.24 brlcad if it's not 0/0/1 on non-remote, then yeah :)
05:10.07 mahesh well, i can see only half the image (that is work done by master node). The other half, the rmeote node should have done it
05:10.28 brlcad should be able to turn on the ray-debug flag and see what the remote node is doing
05:10.38 brlcad how are you setting up the application structure on the remote node?
05:12.24 mahesh all the nodes have access to .g file
05:12.31 mahesh all of them do rt_dirbuild
05:12.53 mahesh then, only master does view_init
05:13.21 brlcad they have access over nfs or you manually copy it?
05:13.35 brlcad or program sends it maybe?
05:13.47 mahesh as of now, i am simulating multiple nodes on single machine
05:13.56 brlcad gotya
05:14.45 mahesh then all of them call do_frame and inturn call do_run
05:15.31 brlcad hmm
05:15.37 mahesh there i call bu_distributed which is the function where nodes get chunks of pixels
05:15.58 brlcad heh, cool
05:16.06 mahesh and then they call mpi_worker
05:16.09 mahesh similar to worker
05:16.30 mahesh master node calls do_pixel
05:16.37 mahesh other nodes return ap structure to master
05:17.05 mahesh and master calls do_pixel on the received ap
05:17.19 mahesh does it all make sense?
05:17.25 brlcad yeah, i think so
05:18.59 brlcad how are you managing the cpu resource structures?
05:19.22 mahesh hmm....haven't touched them at all
05:19.29 brlcad e.g. do_pixel() obtains the application structure from resource[cpu]
05:20.06 brlcad so you would have had to copy it into resource[cpu] before do_pixel at least.. maybe earlier
05:20.27 brlcad that would explain the black pixel "misses"
05:21.32 mahesh remote node calls the same do_pixel function
05:21.44 mahesh so is it not take care of
05:21.48 mahesh may be i am missing something here
05:23.03 brlcad as the rays are fired, the application structure is set with the details of each ray being fired
05:23.31 brlcad that's what a_ray is in the structure
05:24.56 brlcad so when it fires a run of pixels in do_run(), it ends up calling worker()
05:25.11 mahesh right
05:25.18 brlcad worker is like do_pixel(), it assumes that the application structure was stashed into resource[cpu]
05:25.26 brlcad so that it knows where to shoot the ray
05:25.33 mahesh oh i see
05:25.42 brlcad it sounds like your remote nodes are actually shooting a ray successfully
05:25.55 brlcad they are just shooting at "nothing" since they were never told in which direction
05:26.03 mahesh oh i get it
05:27.14 mahesh any clue what field needs to be fixed in my case?
05:27.15 *** join/#brlcad DTRemenak (n=DTRemena@c-67-187-164-66.hsd1.ca.comcast.net)
05:27.21 brlcad for each pixel, a_ray is updated with the new ray direction and worker() does it's magic (calling do_pixel())
05:27.54 brlcad you said you have master calling view_init()?
05:28.01 mahesh yeah
05:28.16 brlcad what do the remote nodes do? how do they start up?
05:29.15 mahesh at the end of do_pixel, master node does a view_pixel where as other nodes return application structure
05:29.39 brlcad that's what the master does.. what do the remote ones do? :)
05:29.40 mahesh in worker function, i store all these ap structures
05:30.05 mahesh remote nodes return ap structure to worker function
05:30.16 mahesh worker function stores them all in an array
05:30.58 mahesh once worker function is complete, the remote node send the array of ap structures back to master
05:31.51 brlcad heh, back up
05:32.03 brlcad to the beginning.. before they return their ap structure
05:32.10 mahesh ok
05:32.13 brlcad what do they do from the beginning?
05:32.41 mahesh shall i send you the worker.c file?
05:33.06 brlcad eventually, but right now, just the general overview :)
05:33.21 mahesh i thought we could both go over
05:33.27 pra5ad my back..
05:33.28 brlcad ok
05:34.42 mahesh sent
05:35.21 brlcad in general, I believe what you're going to want to do with what you have is to have the master node send his application structure to the remote nodes (or maybe even just resource[cpu] where "cpu" is each remote node..not sure if that'll work) or maybe just the ray info
05:35.44 brlcad and each remote node will take that ap structure or ray info and shoot the ray(s)
05:36.24 mahesh hmm
05:36.36 brlcad got it
05:37.16 mahesh i can start from do_run or ?
05:40.45 brlcad depends
05:40.52 brlcad is this a single-processor machine?
05:40.57 mahesh yeah
05:41.23 brlcad and brl-cad was compiled non-parallel presumably..
05:41.39 brlcad i.e. your bu_distributed() in do_run() is actually getting called
05:41.47 mahesh yeah
05:45.28 brlcad see in do_pixel() where it calls rt_shootray(a)?
05:45.51 brlcad that's where it fires the actual ray, passing the application structure that presumably was already filled in with the ray details
05:46.14 brlcad do_pixel() presumes that resource[cpu] was already set to your application structure
05:47.26 brlcad heh, yikes: struct application app[300000];
05:47.49 mahesh i know it is ugly :)
05:48.11 mahesh i just wanted to get all of them working....so...
05:48.45 mahesh so, you think resource[cpu] is not filled?
05:49.00 brlcad not, positive
05:49.11 brlcad actually, could simply be that application is never initialized
05:49.18 brlcad i don't see that anywhere
05:49.50 brlcad when rt starts, there is a global application structure that it uses/initializes
05:49.58 brlcad the parent/master node initialized it
05:49.59 mahesh right
05:50.14 brlcad the parent needs to get that ap to the client to set his global
05:50.29 mahesh where does the master initialize it?
05:51.03 brlcad it's continually initialized as the raytrace proceeds
05:51.47 brlcad if anything, should wait until after grid_setup() as that does initialize the ap structure
05:51.58 brlcad which is in do_pixel ..
05:52.11 brlcad so before bu_distributed, or inside bu_distributed
05:53.02 brlcad yeah.. inside bu_distributed
05:53.19 brlcad master should set ap to client, client should set his ap to what master sent
05:53.55 mahesh before mpi_worker is called right?
05:55.22 brlcad master sends before mpi_worker(), non-master sends ap structure back
05:55.36 brlcad not just a_color (which would be have been ap.a_color, btw)
05:55.45 mahesh got it
05:56.17 mahesh but how about resource[cpu]?
05:56.48 brlcad forget about it for now :)
05:56.49 Guu` I've never heard of about it for now :), brlcad!
05:57.00 brlcad you might even get it for free
05:57.35 brlcad that makes sense now
05:58.24 brlcad the local master has the ap structure, so his rt_shootray() is using the initialized one.. the remote nodes, however, are not so it's using them uninitialized and missing
05:59.05 brlcad you're making great progress! :)
05:59.19 mahesh making fun of me :)
05:59.26 brlcad no seriously
05:59.28 brlcad it's great
05:59.46 mahesh thanks!
05:59.54 brlcad you've had other stuff going on, i'm sure too
06:00.02 mahesh yeah
06:00.14 brlcad awesome to see code for it, and it's making sense so far
06:01.20 mahesh once i see a full image on the screen, i will start playing with some ideas i have
06:01.41 brlcad have you modified other files?
06:01.42 mahesh hopefully, the code will be much more clean and robust
06:01.45 mahesh yeah
06:02.18 mahesh functions in view.c are called only my master
06:02.23 mahesh i have changed that
06:02.35 mahesh and small changes in do.c
06:02.56 mahesh i think those are the only files i have changed (plus ofcourse main.c)
06:33.04 brlcad well, i'm excited to hear that you've gotten it to render - keep an eye on transferring the ap structure
06:33.07 brlcad you can "set" it once it's sent mby simply performing a struct copy
06:33.26 brlcad or there are examples throughout the code that do the struct copy too, it's a one-liner
06:34.40 brlcad alas, for now, I must zzz.. haven't in over 24 hours
06:36.30 mahesh sure
07:45.18 *** join/#brlcad PKMOBILE (n=Apathy@c-68-33-243-45.hsd1.md.comcast.net)
11:22.27 *** join/#brlcad _AchiestDragon (n=dave@whipy.demon.co.uk)
11:22.50 *** join/#brlcad boemann (n=cbr@213.237.55.84.adsl.nivaa.worldonline.dk)
11:23.18 boemann hi
11:23.34 boemann what kind of widget toolkit does brlcad use?
11:25.02 boemann and does a .deb exist?
11:47.39 *** part/#brlcad AchiestDragon (n=dave@whipy.demon.co.uk)
11:59.11 *** join/#brlcad archivist_ (n=archivis@host217-35-76-52.in-addr.btopenworld.com)
12:00.50 ``Erik libtk and not that I've seen
12:09.02 boemann ok thanks
12:11.48 ``Erik tho someone was talking about making a deb not too long ago
12:12.16 boemann it is listed as being worked on for the last 365 days :(
12:17.07 boemann what is the shape of the code - i mean does it show it's age - is it object oriented?
12:17.56 boemann mainly regarding rt and db
12:19.46 ``Erik uh, it's C, rt is well encapsulated and kept up to date, and I kinda think it's better than most OO stuff
12:20.21 ``Erik ancient cruft is culled on occasion, like all the k&r style is gone in favor of ansi style... o.O
12:20.33 boemann well C doesn't mean it can't be OO ;)
12:20.56 boemann in design that is
12:20.59 ``Erik true, but then you leave the world of tightly defined oo, since C affords you far more flexability
12:21.19 ``Erik the core of librt, as far as I understand, is a big honkin' struct with a couple callbacks in it
12:21.37 boemann ok
12:22.03 boemann how about opengl accel of the editor
12:22.07 ``Erik but I have only a passing knowledge of those things... most of my involvement was rewriting the build system to use automake, so 99% of my job was 'does it compile?'
12:22.19 boemann :)
12:22.43 ``Erik yeah, it has that ability, provided the appropriate display manager is used and the X you're using has ogl all ready to go
12:22.58 boemann good
12:22.59 ``Erik heh, it might even still support irisgl, too :D
12:23.12 boemann :)
12:23.31 boemann i did some iris gl once
12:25.52 boemann I'm considering joining in - though I'm not sure I wouldn't rather start from scratch with C++
12:26.39 boemann perhaps only make a C++ wrapper for the brl core
12:26.53 boemann and then make my own editor
12:28.33 boemann does brl support concurrent development from different workstations?
12:29.09 boemann not development but work
12:33.52 ``Erik um, multiple participants having the .g file open, you mean?
12:34.48 ``Erik I BELIEVE the current technique for having several people working on the same model at the same time involves having the model broken into several pieces, the people work on their own piece, then they glue the pieces together at the end into the final output file
12:35.03 ``Erik but I could be way off base with that, best to talk to brlcad when he wakes up
12:40.18 boemann where do the majority of developers live?
12:40.51 ``Erik uhm, maryland, usa, I'd imagine
12:41.05 ``Erik but there are listed developers and many contributors all over the world o.O
12:41.22 ``Erik iirc, there's been a recent surge of activity from someone in italy o.O
12:42.10 ``Erik but I think most developers work for arl, formerly known as brl... :)
12:43.11 boemann so there is real work going on
12:43.33 ``Erik yeah
12:43.38 ``Erik well
12:43.49 ``Erik sean is supposed to be working on it... (he goes by brlcad here...)
12:44.02 boemann nice to know - i don't want to involve myself in something dead
12:44.21 ``Erik lee does stuff on occasion, and when I get the excuse/time, I make some fixes for the bsd's
12:45.25 ``Erik I have to go to work, I'll be back in an hour if you have more questions... later :)
12:45.37 boemann ok bye
12:45.44 boemann i wont be here
12:46.14 boemann thanks for all the help
12:46.24 *** part/#brlcad boemann (n=cbr@213.237.55.84.adsl.nivaa.worldonline.dk)
18:01.00 *** join/#brlcad clock_ (n=clock@84-72-62-232.dclient.hispeed.ch)
18:01.35 CIA-5 BRL-CAD: 03bob1961 * 10brlcad/src/tclscripts/swidgets/scripts/tree.itk: Check-in for Doug Howard. Added a -sortmode option.
18:02.12 clock_ brlcad: hi, I learnt riding buttonlift today :)
18:19.20 brlcad ~buttonlift
18:19.46 brlcad ah, heh, neat
18:26.42 ``Erik skiing?
18:47.22 clock_ ``Erik: snowboarding
18:47.38 clock_ brlcad: I completed it 11 times! :)
18:50.29 clock_ brlcad: are you going bodybuilding?
18:50.57 brlcad working on it
18:51.05 brlcad competition in a couple weeks
18:51.47 clock_ brlcad: eee? Can I see your pictures?
18:51.56 brlcad heh, I don't have any pics
18:52.25 clock_ brlcad: what kind of competition?
18:52.26 brlcad i'd need to trim down some more for pics
18:52.35 brlcad lifting competition, bench
18:52.54 clock_ brlcad: hey this is interesting. Those projects whose authors do sports tend to not have problems understanding their users
18:53.20 brlcad heh
18:53.39 clock_ like OpenBSD is cool they have good user manual, funny cartoons, first fixing bugs and only then doing features, caring about security, and Theo de Raadt is said to be always in the mountains :)
18:54.02 clock_ brlcad: is it possible that iron turns into wood because of cosmic rays bombarding the atomic kernels inside?
18:54.06 ``Erik theo's also a major asshole :D but on occasion, he's not TOTALLY full of shit...
18:54.11 ``Erik at least, that's what I've heard
18:54.40 brlcad i'm going to have to trim back on the lifting though, going to join the rowing team
18:54.51 clock_ brlcad: I started with iron dumbell 10kg and it started feeling like wooden so I had to add some plates so now it reads 12.5kg but soon I got the same syndrome again
18:54.59 ``Erik what rowing team?
18:55.06 clock_ ``Erik: what does major asshole mean?
18:55.09 brlcad baltimore has a team
18:55.21 ``Erik funky
18:55.28 clock_ actually what does asshole mean? I know it's the hole into the rectum but can't imagine which personality traits it's supposed to imply.
18:55.41 ``Erik clock: an unpleasant person...
18:55.51 brlcad clock_: what's your native language? :)
18:55.58 clock_ brlcad: czech
18:56.05 clock_ hehe :)
18:56.16 clock_ brlcad: is it national competition where youre going?
18:56.35 ``Erik translation table? o.O
18:56.46 brlcad clock_: maybe that he's a "hovnohlava"? :)
18:57.15 brlcad usually implies a bit of pretencious impatience
18:57.50 brlcad ``Erik: multilingual swear list filtering system I wrote for bz has several dozen languages, swears and insults out the wazoo
18:58.15 clock_ brlcad: hovnohlava doesn't mean anything in czech
18:58.48 clock_ brlcad: we say "prijemnej jak osina v prdeli" - as pleasant as an wan in the ass
18:58.53 clock_ wan -> awn
18:59.20 brlcad clock_: maybe you could update my czech section then :)
18:59.24 brlcad http://cvs.sourceforge.net/viewcvs.py/bzflag/bzflag/misc/multilingualSwearList.txt?rev=HEAD
18:59.37 ``Erik what about "pretentious arrogant bastard"?
18:59.50 clock_ ``Erik: pretentious means that he's pretending things?
18:59.59 brlcad can fill it in with whatever swear words you can think of :) .. but has to be not just rude but blatently insulting
19:00.36 ``Erik http://m-w.com/dictionary/pretentious
19:00.41 clock_ brlcad: what kind of competition are you engaging in? Are you a serious bodybuilder?
19:00.55 clock_ brlcad: trim down == make the muscles more cut?
19:01.03 ``Erik http://www.urbandictionary.com/define.php?term=asshole
19:01.39 clock_ pretentive impatient unpleasant person that's exactly me :)
19:01.41 ``Erik hehehe, #2 :D do you know a jenna tools, sean?
19:02.06 clock_ No, I won't update the filtering list, because that way I would work against myself ;-)
19:02.25 clock_ but I am getting better in getting rid of these personality traits
19:02.48 clock_ and I am erasing them in the brain and filling the gained space with skills like snowboarding and skateboarding ;-)
19:03.17 ``Erik and running into trees and rocks and stuff are helping to knock them clean out of your skull? ;) *duck*
19:03.46 clock_ ``Erik: I smashed my head once when they didn't have a helmet available in the renting - after that I bought my own helmet :)
19:03.47 ``Erik adolph hitler is a swear?
19:04.15 clock_ ``Erik: adolph hilter isn't anymore, he's dead
19:04.23 ``Erik "bigbooty"? ... O.o
19:04.43 ``Erik hey, what's wrong with buttface?
19:04.44 ``Erik hehehe
19:04.46 clock_ ``Erik: hey I saw shaun white
19:05.24 clock_ ``Erik: what are you doing there in the USA so that you have so cute snowboarders?
19:05.41 ``Erik they drink milk
19:05.48 ``Erik it does a body good o.O
19:06.41 ``Erik let us know how those tits fill out.
19:06.57 brlcad jenna tools?
19:06.59 brlcad nope
19:07.31 brlcad clock_: trim down means get rid of some of my fat ;)
19:08.10 clock_ brlcad: some news on brlcad? An integrated "warhead" command with syntax warhead <x> <y> <z> <yield> <length> that makes nuclear warhead design macro?
19:08.19 brlcad clock_: aww, updating that list could be your contribution to the "clean" servers that try to impose their morality on other players :)
19:08.33 brlcad that'd be just a few lines of tcl :)
19:09.09 clock_ brlcad: how are you trimming down? Eating lots of little meals?
19:09.26 brlcad drinking less beer
19:09.32 clock_ hehe :)
19:09.38 brlcad :)
19:09.45 clock_ so it is probably not a world bodybuilding championship :)
19:09.49 brlcad though today was a regression
19:10.05 brlcad nah, not a major event, local event
19:10.15 brlcad have to start somewhere
19:11.14 brlcad ``Erik: surprisingly enough, that name is often more inciteful than most swears on some game servers
19:11.36 brlcad some of the words in the massive english section could be removed
19:12.02 brlcad the list was compiled from lots of sources and cleaned up
19:12.54 ``Erik amusing reading :)
20:23.41 CIA-5 BRL-CAD: 03brlcad * 10brlcad/src/mged/titles.c: mild sanity check, even if the directory pointer is valid, make sure the name is too before printfing
20:24.32 clock_ "mild sanity" that's a beautiful term ;-)
20:28.48 CIA-5 BRL-CAD: 03brlcad * 10brlcad/src/mged/ged.c: don't attempt to call do_titles() with an overlay_vls that hasn't been initialized. this 'also' fixes sf bug 1326045 that was fixed by another edit as well
20:30.40 brlcad heh
20:31.02 clock_ bug 1326045?
20:31.25 brlcad yes?
20:31.32 brlcad sf bug tracker ids
20:31.48 clock_ if the price of brlcad was increased 1 dollar for every bug fixed then I am already a millionaire! :)
20:31.49 brlcad your bug submission :)
20:32.31 clock_ brlcad: my muscles get bigger from lifting weight, cannot this indicate some disease? :)
20:34.13 brlcad a healty disease ;)
20:34.53 clock_ brlcad: lifting weights and then going in front of a mirror is the cheapest way to get gay porn :)
20:35.41 brlcad heh, I've never quite heard it put that way before
20:35.49 brlcad self-indulgence
20:35.51 clock_ or like the one with a god:
20:36.04 clock_ can a bodybuilder's biceps grow so big that he cannot carry it anymore?
20:36.24 brlcad probably :)
20:37.11 clock_ brlcad: on Sunday I exceeded the critical mass of sport from which there is no return back ;-)
20:37.51 clock_ brlcad: it's like a drug when I don't feel delayed onset muscle soreness my hands shake and I am nervous and I have to go doing some sport
20:38.48 brlcad how about this: http://www.t-nation.com/findArticle.do?article=211hate
20:41.02 clock_ brlcad: what's synthol? car brake fluid?
20:42.52 brlcad steroid, i think
20:47.55 ``Erik hah
20:48.26 clock_ brlcad: that's 100% great :)
20:50.19 ``Erik "posing oil" o.O
20:50.57 ``Erik nasty, it's an oil compound that settles between muscle fibers, increasing size (but not strength)
20:58.49 clock_ brlcad: are you bodybuilding to attract women or do you already have a wife?
20:59.30 clock_ I wonder how the guy can bend his arms despite those basketballs
20:59.40 clock_ it's like boobs on the wrong place
21:00.47 brlcad clock_: I just workout to feel good, I enjoy it
21:04.43 clock_ brlcad = mental workout ;-)
21:04.49 clock_ brlcad: does your brain gain inches? :)
21:06.49 brlcad i read research papers in my spare time, but I think my head leaks .. can't remember ;)
21:11.25 ``Erik you have spare time? how'd you pull that one off?
21:13.46 brlcad during commercials between episodes of stargate and bsg
21:14.57 ``Erik heh
21:22.03 clock_ brlcad: put your muscle photo on brlcad page with label "these muscles are very solidly modelled"
21:24.30 brlcad heh
22:38.15 *** join/#brlcad cad790 (n=3b5c60d0@bz.bzflag.bz)
22:41.57 *** join/#brlcad DTRemenak (n=DTRemena@c-67-187-164-66.hsd1.ca.comcast.net)
23:54.20 ``Erik hrm

Generated by irclog2html.pl by Jeff Waugh - find it at freshmeat.net! Modified by Tim Riker to work with blootbot logs, split per channel, etc.